Spatial Filter Update via Error Matrix Precoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for updating spatial filters in MIMO radio networks incur significant signaling overhead, particularly when dealing with multiple users, as they require exchanging MSE-matrices and channel information among nodes, which becomes cumbersome with a large number of users.
Innovation Solution
The method involves transmitting reference signals precoded by spatial filters and error matrices from transmit nodes to receive nodes, allowing each node to recompute and update its filters independently without additional control signaling, thereby reducing signaling overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If spatial filters are updated using conventional methods with MSE-matrix exchange, then filtering accuracy is improved, but signaling overhead increases significantly
Solution Approach 1:
The patent extracts only the essential information needed for spatial filter updating by having receive nodes transmit condensed reference signals containing error matrix information, rather than exchanging complete MSE-matrices. This extraction principle reduces the amount of data that must be signaled while preserving the critical information needed for filter optimization.
Solution Approach 2:
The patent inverts the conventional approach by having receive nodes actively transmit reference signals with error matrix information back to transmit nodes, rather than having transmit nodes passively receive MSE-matrices. This inversion enables transmit nodes to independently recompute filters without relying on direct MSE-matrix exchanges, thereby reducing signaling overhead.
2Productivity
If spatial filters are updated with complete MSE-matrix information, then filter optimization is improved, but system complexity increases
Solution Approach 1:
The patent applies local quality by enabling each transmit node to independently recompute error matrices and update its own spatial filters using locally available reference signals from receive nodes. This distributed local computation eliminates the need for centralized MSE-matrix management and reduces overall system complexity while maintaining optimization efficiency.
3Productivity
If reference signals are transmitted with error matrix precoding, then spatial filter updating efficiency is improved, but signal processing complexity increases
Solution Approach 1:
The patent applies preliminary action by having receive nodes pre-compute error matrices based on current spatial filters and channel conditions, then embed this information into reference signals before transmission. This preliminary preparation enables transmit nodes to directly reuse the error matrix information for filter updates without performing complex computations, thereby improving updating efficiency while managing processing complexity at the receive end.
Data Source
AI summary
A method of updating spatial filters in a radio network comprising at least two transmit nodes each in radio communication with at least one receive node on a Multiple-Input Multiple-Output (MIMO) radio channel, comprises transmitting, from the respective transmit node, first reference signals precoded by a first spatial filter of the respective transmit node; and receiving, from the receive nodes, second reference signals that are precoded using a second spatial filter of the respective receive node and an error matrix of the respective receive node. The second spatial filter depends on a channel estimate based on the transmitted first reference signals. The error matrix is indicative of an error of the first and second spatial filters in equalizing the MIMO radio channel. The method further comprises recomputing, for each of the at least one receive node in radio communication with the respective transmit node, the error matrix of the respective receive node; and updating the first spatial filter of the respective transmit node using the recomputed error matrix.